Method and device for predicting mobility and adsorbability of shale oil based on pore structure parameters

Abstract

The invention belongs to the technical field of petroleum geological exploration, and particularly relates to a method and device for predicting mobility and adsorbability of shale oil based on pore structure parameters. The method comprises the following steps that S1, mineralogy and basic geochemical parameters are used, Misforest iterative interpolation is adopted to supplement pore structure parameters, and pore structure parameter data are obtained; s2, the mineralogy type, the basic geochemical parameters, the porosity, the pore structure parameter data and the multi-temperature-order pyrolysis data form an original data set; and S3, inputting the original data set into the prediction model to obtain a prediction result of the mobility and the adsorbability of the shale oil. According to the method, in order to realize accurate prediction of the OMC and the OSC, pore structure parameters are supplemented by adopting a MissForest iterative interpolation method, and the OMC and the OSC of the shale are calculated based on a machine learning model.

Description

Technical Field

[0001] The invention belongs to the technical field of petroleum geological exploration, and specifically relates to a method and a device for predicting the mobility and adsorption of shale oil based on pore structure parameters. Background Art

[0002] With the rapid development of development technologies such as horizontal drilling and multi-stage hydraulic fracturing, shale oil and gas are currently the focus of exploration and development in my country. Taking the United States as an example, shale oil accounted for more than 40% of oil and gas production in 2020. In addition, in recent years, my country has also achieved many remarkable results in the exploration and development of continental shale oil and gas. However, although the current continental shale oil and gas exploration results are remarkable, due to the strong heterogeneity caused by the continental sedimentary background, the optimization of continental shale oil and gas sweet spots is still a challenging task.

[0003] Previous studies on continental shales have demonstrated the important influence of oil mobility on the development of continental shale reservoirs. There are two states of occurrence of oil in shale, free state and adsorbed state, and there are obvious differences in the mobility of oil in different states. Previous studies have shown that the mobility of oil in shale is affected by many factors, including thermal maturity, kerogen type, mineral composition, and geological parameters such as pore structure. At present, the understanding of the controlling factors of oil mobility in shale is still unclear, which makes it difficult to predict oil mobility.

[0004] In order to characterize the content of oil in different states of occurrence in shale, a variety of methods have been developed, including: multi-stage pyrolysis, step-by-step extraction, two-dimensional nuclear magnetic resonance and other methods. Different methods have their own unique applicability and limitations. Among them, multi-stage pyrolysis is one of the most important methods for evaluating the mobility of oil in shale due to its low cost and accurate results. However, compared with conventional pyrolysis tests, the cost of multi-stage pyrolysis is still relatively high, so there is a general lack of sufficient relevant data in the industry, which in turn restricts shale oil and gas exploration and development and the understanding of the mobility of shale.

[0005] Studies have shown that pore structure parameters have an important influence on the mobility and adsorption of shale oil. However, due to the high cost of analyzing and testing pore structure parameters of shale oil, the industry generally lacks a large amount of pore structure data, which undoubtedly restricts the evaluation of shale reservoirs and our understanding of the mobility of shale oil. Interpolation of a large amount of missing pore structure data is a common processing method, but simple interpolation will bring huge uncertainty. Stekhoven and Bühlmann (2012) proposed an iterative interpolation method based on random forest (Missforest), which has been well applied in data processing in the field of bioinformatics engineering, but has not yet received widespread attention in the field of unconventional oil and gas exploration, which restricts researchers from making rational use of the analysis and test results with a large amount of missing data (samples are difficult to obtain and the test cost is expensive).

[0006] Machine learning (ML) has become a powerful tool for mining complex patterns in data. However, ML is essentially a black box model, and it is difficult to understand how the model makes predictions by understanding its internal mechanisms. To overcome this challenge, interpretable machine learning techniques have received widespread attention. These techniques can quantify the importance of each input feature to model predictions from different perspectives, and can also provide insights into the complex synergies between input features. Summary of the invention

[0007] In order to accurately and cost-effectively analyze the mobility and adsorption of shale oil, the present invention provides a method for iteratively interpolating pore structure parameters using Missforest and predicting the mobility and adsorption of shale oil based on interpretable machine learning.

[0008] In order to achieve the above-mentioned object of the invention, the present invention provides the following technical solutions:

[0009] The method for predicting the mobility and adsorption of shale oil based on pore structure parameters comprises the following steps:

[0010] S1, using mineralogy and basic geochemical parameters, the pore structure parameters are supplemented by Missforest iterative interpolation to obtain pore structure parameter data;

[0011] S2, the original data set is composed of mineralogy type, basic geochemical parameters, porosity, pore structure parameter data and multi-temperature pyrolysis data;

[0012] S3, inputting the original data set into the prediction model to obtain the prediction results of shale oil mobility and adsorption.

[0013] As a preferred embodiment, the mineralogical types include quartz, feldspar, plagioclase, potassium feldspar, calcite, dolomite, aragonite, siderite, gypsum, anhydrite, clay minerals and pyrite; the basic geochemical parameters include organic matter abundance and pyrolysis hydrocarbon peak temperature.

[0014] As a preferred solution, the pore structure parameter data include volumes and specific surface areas of different types of pores, wherein the pore types include micropores, transitional pores, mesopores and macropores.

[0015] As a preferred solution, during the training of the prediction model, the calculation formula for the mobility and adsorption of shale oil is:

[0016]

[0017] Among them, OMC is the mobility of shale oil, OSC is the adsorption of shale oil, TOC is the abundance of organic matter, S 1-1 It is a multi-stage pyrolysis light oil component, S 1-2 It is a multi-stage pyrolysis light medium oil component, S 2-1 It is a multi-temperature stage pyrolysis of heavy hydrocarbons and colloidal asphalt components.

[0018] As a preferred solution, during the prediction model training process, the regression model is used for evaluation, and the coefficient of determination and root mean square error used in the evaluation are defined as follows:

[0019]

[0020] In the formula, is the predicted data, y i is the actual value, is the actual data mean, n is the number of each data set, R 2 is the coefficient of determination and RMSE is the root mean square error.

[0021] As a preferred solution, the prediction models include MLP-SGD, MLP-Lbfgs, MLP-Adam, RF-PSO, RF-GA, RF-DE, XGBoost-PSO, XGBoost-GA, XGBoost-DE, SVM-PSO, SVM-GA, and SVM-DE.

[0022] As a preferred solution, the preferred models in the prediction model include RF-PSO and RF-PSO.

[0023] As a preferred solution, step S4 is also included: interpreting the prediction model using a variety of interpretability algorithms to obtain the main influencing factors of shale oil mobility and adsorption.

[0024] As a preferred solution, the explainability algorithms include PFI, SHAP, and PDP.

[0025] Based on the same concept, a device for predicting the mobility and adsorption of shale oil based on pore structure parameters is also proposed, comprising at least one processor and a memory communicatively connected to the at least one processor; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute any of the above-mentioned methods for predicting the mobility and adsorption of shale oil based on pore structure parameters.

[0026] Compared with the prior art, the present invention has the following beneficial effects: In order to achieve accurate prediction of OMC and OSC, the present invention aims at the current situation that pore structure parameter data are relatively scarce and generally missing in the process of petroleum geological research. The Misssforest iterative interpolation method is adopted to supplement the pore structure parameters with mineralogy, basic geochemistry and porosity. Subsequently, based on the mineralogy, basic geochemistry and pore structure data, a variety of machine learning-based models are constructed and compared to screen out the optimal OMC and OSC prediction models, thereby accurately calculating the OMC and OSC of shale. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 is a flow chart of the method for predicting shale oil mobility and adsorption based on pore structure parameters in Example 1;

[0028] Figure 2 This is a flow chart of an improved shale oil OMC and OSC prediction method in Example 2;

[0029] Figure 3 is the correlation feature between the feature and the target in Example 2;

[0030] Figure 4 is the optimal result obtained by the 12 machine learning models in Example 2;

[0031] Figure 5 is a cross plot of the predicted values ​​and actual values ​​of the optimal OMC and OSC prediction models in Example 2;

[0032] Figure 6 is the optimal OMC and OSC prediction model replacement importance ranking diagram in Example 2;

[0033] Figure 7 It is the SHAP ranking diagram of the optimal OMC and OSC prediction model in Example 2;

[0034] Figure 8 It is the result of the main characteristics of the optimal OMC and OSC prediction model in Example 2 that partly depends on the results;

[0035] Fig. 9This is a performance comparison diagram of the OMC and OSC models in Example 2 with and without considering the pore structure parameters. DETAILED DESCRIPTION

[0036] The present invention is further described in detail below in conjunction with test examples and specific implementation methods. However, this should not be understood as the scope of the above subject matter of the present invention being limited to the following embodiments, and all technologies realized based on the content of the present invention belong to the scope of the present invention.

[0037] Example 1

[0038] The present invention provides a method for iteratively interpolating pore structure parameters using Missforest, and predicting shale oil mobility and adsorption based on interpretable machine learning. First, the missing pore structure parameter data are iteratively interpolated using mineralogy and basic geochemical parameters. Subsequently, the mobility (OMC) and adsorption (OSC) of shale oil are rapidly predicted through mineralogy, basic geochemistry and pore structure parameter data, which is fast and efficient. The prediction model is analyzed using a variety of interpretable techniques to reveal the influence mechanism of mineral components, geochemical parameters and pore structure parameters on the mobility and adsorption of shale oil.

[0039] The method for predicting the mobility and adsorption capacity of shale oil based on pore structure parameters is shown in the flow chart below. Figure 1 As shown, the following steps are included:

[0040] S1, using mineralogy and basic geochemical parameters, the pore structure parameters are supplemented by Missforest iterative interpolation to obtain pore structure parameter data;

[0041] S2, the original data set is composed of mineralogy type, basic geochemical parameters, porosity, pore structure parameter data and multi-temperature pyrolysis data;

[0042] S3, inputting the original data set into the prediction model to obtain the prediction results of shale oil mobility and adsorption.

[0043] As a preferred solution, after obtaining the prediction results of shale oil mobility and adsorption, S4 is also included, in which the optimal model is explained using a variety of interpretability algorithms to obtain the main influencing factors of shale oil mobility and adsorption.

[0044] Specifically, the method for predicting the mobility and adsorption of shale oil based on interpretable machine learning is as follows:

[0045] Step (1) collects shale mineralogy (quartz, feldspar, plagioclase, potassium feldspar, calcite, dolomite, aragonite, siderite, gypsum, anhydrite, clay minerals and pyrite), basic geochemistry (TOC, Tmax), porosity, pore structure parameters (including volume PV and specific surface area SSA of different types of pores, where pore types include micropores, transitional pores, mesopores and macropores) and multi-temperature pyrolysis data (S 1-1 , S 1-2 , S 2-1 , S 2-2 ), as the original data set;

[0046] Step (2) performs preprocessing operations on the raw data in step (1), performs outlier detection and missing value mean interpolation on the raw data. Since the acquisition cost of pore structure parameter data is significantly higher than other tests, the Missforest iterative interpolation method is used to supplement the missing pore structure parameters with mineralogy and basic geochemical parameter data.

[0047] Step (3), calculate OMC and OSC based on TOC and distribution pyrolysis data, the calculation formula is as follows:

[0048]

[0049] Among them, OMC is the mobility of shale oil, OSC is the adsorption of shale oil, TOC is the abundance of organic matter, S 1-1 It is a multi-stage pyrolysis light oil component, S 1-2 It is a multi-stage pyrolysis light medium oil component, S 2-1 It is a multi-temperature stage pyrolysis of heavy hydrocarbons and colloidal asphalt components.

[0050] Step (4), dividing the preprocessed data into two parts: features and targets. The features include mineralogy, basic geochemical parameters, and pore structure parameters. The targets include OMC and OSC. The original data are divided into two subsets according to different targets.

[0051] Step (5), before model development, the data subsets are randomly grouped into training sets and test sets. The training set accounts for 80% of the data subset and is used for ML model development; the test set accounts for 20% of the data subset and is used to evaluate the generalization ability of the optimal model. In order to reduce the impact of the dimensions of different features on model prediction, after dividing the training set and the test set, the training set is standardized, and the test set is standardized using the mean and standard deviation obtained from the training set to avoid data leakage caused by introducing test set data during the model development phase;

[0052] Step (6), in the training set data, four classic machine learning algorithms (SVM, RF, XGBoost, MLP) are used to build the model, and six optimization algorithms (PSO, GA, DE, SGD, Lbfgs, Adam) are used in combination, and finally 12 machine learning models (MLP-SGD, MLP-Lbfgs, MLP-Adam, RF-PSO, RF-GA, RF-DE, XGBoost-PSO, XGBoost-GA, XGBoost-DE, SVM-PSO, SVM-GA, SVM-DE) are developed. It should be emphasized that SGD, Lbfgs and Adam are used to optimize weights and biases in the training phase of the MLP model. DE, GA and PSO algorithms are responsible for optimizing the hyperparameters in the RF, XGBoost and SVM models.

[0053] In step (7), common metrics of regression models, such as the coefficient of determination (R2) and root mean square error (RMSE), are applied during model development and evaluation, as defined below:

[0054]

[0055] In the formula, is the predicted data, y i is the actual value, is the actual data mean, and n is the size of each data set. The average prediction performance based on five-fold cross validation (R2CV and RMSECV) was used to evaluate 12 machine learning models and then select the optimal model;

[0056] Step (8), use PFI, SHAP, and PDP techniques to interpret the optimal OMC and OSC prediction models obtained in step (6).

[0057] The abbreviations and key terms are defined as follows: S1: free hydrocarbon content by conventional pyrolysis; S2: cracked hydrocarbon content by conventional pyrolysis; S1-1: light oil component by multi-temperature-stage pyrolysis; S1-2: light and medium oil component by multi-temperature-stage pyrolysis; S2-1: heavy hydrocarbon and colloid asphalt component by multi-temperature-stage pyrolysis; S2-2: regenerated hydrocarbon by pyrolysis cracking of kerogen by multi-temperature-stage pyrolysis; Tmax: peak temperature of pyrolysis hydrocarbon; TOC: organic matter abundance; OMC: shale oil mobility; OSC: shale oil adsorption; SVM: support vector machine algorithm; RF: random forest algorithm; XGBoost: extreme gradient boosting tree; MLP: multi-layer perceptron; PSO: particle swarm algorithm; GA: genetic optimization algorithm; DE: differential optimization algorithm; PFI: permutation feature importance; SHAP: Shapley addition interpretation.

[0058] Example 2

[0059] In order to solve the problem of missing research data on shale oil mobility and improve the understanding of the factors controlling shale oil mobility, the present invention proposes a prediction method for shale oil OMC and OSC based on interpretable machine learning using mineral composition, geochemical parameters and pore structure parameters. Figure 2 A flow chart of an improved shale oil OMC and OSC prediction method according to an embodiment of the present invention is shown.

[0060] The research area of ​​the embodiment of the present invention is located in a certain area, and the research conducted a series of analytical tests on black shales from the Qianfoya Formation and the Da'anzhai Section. The research data include TOC test, multi-step pyrolysis test, X-ray diffraction test, helium porosity, nitrogen adsorption, mercury injection capillary pressure, etc.

[0061] According to the above steps, the mineralogy (quartz, feldspar, plagioclase, potassium feldspar, calcite, dolomite, aragonite, siderite, gypsum, anhydrite, clay minerals and pyrite) and basic geochemistry (TOC, T max ), porosity and multi-temperature pyrolysis data (S 1-1 , S 1-2 , S 2-1 , S 2-2 ), as the original data set.

[0062] Subsequently, the raw data were preprocessed, including outlier detection and missing value interpolation. In view of the serious problem of missing pore structure parameters (pore volume and specific surface area of ​​micropores, transitional pores, mesopores and macropores), Missforest iterative interpolation was used to supplement the pore structure parameters. Table 1 shows the results of Missforest iterative interpolation. The higher the crossover error (R 2 CV ) and the out-of-bag sample score (Oob score) confirm the feasibility of the method.

[0063] Table 1 Missforest iterative interpolation performance evaluation

[0064]

[0065] In addition, OMC and OSC were calculated based on TOC and distribution pyrolysis data, and the calculation formula is as follows:

[0066]

[0067] The preprocessed data is divided into two parts: features and targets. Features include mineralogy, basic geochemical parameters, and porosity, and targets include OMC and OSC. The raw data is divided into two subsets according to different targets. Finally, the preprocessed raw data is divided into two subsets: (1) Data subset 1 contains 21 features and corresponding OMCs; (2) Data subset 2 contains 21 features and corresponding OSCs.

[0068] Figure 3 The correlation characteristics between the features and the target in the embodiment are shown. Geochemical parameters (TOC, T max ) provided the strongest correlation trend with the target (R>0.6), TOC was positively correlated with OMC and OSC, T max There is a negative correlation with OMC and OSC. There is a certain degree of negative correlation between the quartz and calcite content and OMC and OSC. It is worth noting that except for the micropore volume and specific surface area, which have an insignificant or very weak correlation with OMC and OSC, other pore structure parameters are positively correlated with OMC and OSC. Compared with OMC, the correlation between pore structure parameters and OSC is stronger.

[0069] Before model development, the data subsets were randomly divided into training sets and test sets. The training set accounted for 80% of the data subset and was used for ML model development; the test set accounted for 20% of the data subset and was used to evaluate the generalization ability of the optimal model. In order to reduce the impact of the dimensions of different features on model prediction, after dividing the training set and the test set, the training set was standardized, and the mean and standard deviation obtained from the training set were used to standardize the test set to avoid data leakage caused by introducing test set data during the model development stage.

[0070] In the training set data, four classic machine learning algorithms (SVM, RF, XGBoost, MLP) were used to build the model, and six optimization algorithms (PSO, GA, DE, SGD, Lbfgs, Adam) were used in combination. Finally, 12 machine learning models (MLP-SGD, MLP-Lbfgs, MLP-Adam, RF-PSO, RF-GA, RF-DE, XGBoost-PSO, XGBoost-GA, XGBoost-DE, SVM-PSO, SVM-GA, SVM-DE) were developed. It should be emphasized that SGD, Lbfgs, and Adam are used to optimize weights and biases in the training phase of the MLP model. DE, GA, and PSO algorithms are responsible for optimizing the hyperparameters in the RF, XGBoost, and SVM models.

[0071] The commonly used evaluation index of regression model, namely the coefficient of determination (R2 ) and root mean square error (RMSE), defined as follows:

[0072]

[0073] In the formula, is the predicted data, y i is the actual value, is the actual data mean, and n is the size of each data set. The average prediction performance (R 2 CV and RMSE CV ) was used to evaluate 12 machine learning models and then select the optimal model.

[0074] Figure 4 The best results obtained by 12 machine learning models are shown. Among them, RF-PSO achieved the best OMC prediction results, with R 2 0.87 and 0.86 respectively, and the RMSE is 0.23 and 0.22 respectively. RF-PSO achieved the best OMC prediction results, with R 2 They are 0.96 and 0.81 respectively, and the RMSE are 0.08 and 0.19 respectively.

[0075] Figure 5 The cross plots of the predicted values ​​and actual values ​​of the optimal OMC and OSC prediction models are shown. It can be observed that the model has a good prediction effect on OMC and OSC, and the predicted OMC and OSC are very similar to the actual values.

[0076] After the optimal model is determined in the model development process, PFI, SHAP, and PDP technologies are used to explain the optimal OMC and OSC prediction models. PFI, SHAP, and PDP are three different interpretable technologies to analyze how the optimal prediction model is output and what factors are more important to the results. Machine learning is a black box model, so it is difficult to really derive its operating mechanism. Therefore, as a preferred solution, multiple interpretable technologies are also used to explain the optimal OMC and OSC prediction models, and then their results are comprehensively analyzed to find out which factors are more important to the results.

[0077] The permutation feature importance graph obtained by PFI, the SHAP global dependency graph obtained by SHAP, and the partial dependency graph obtained by PDP can be combined to provide insights into the controlling factors of OMC and OSC. maxTwo geochemical parameters have the greatest contribution to the accurate prediction of OMC and OSC. In addition, siderite, macropore volume, calcite and pyrite are more important for the prediction of OMC; pyrite, clay minerals, calcite and quartz are more important for the prediction of OSC. Figure 6 The permutation importance ranking of the optimal OMC and OSC prediction models is shown.

[0078] The SHAP value can be used to study the impact of each feature on model prediction, thereby explaining the optimal machine learning model. Figure 7 The SHAP value ranking of each feature of the optimal OMC and OSC models is shown respectively. The features along the y-axis from top to bottom reflect the difference in the SHAP importance of the model. For the OMC model, the first six features of the SHAP summary graph are TOC, T max , calcite, macropore volume, pyrite and porosity; the first six features of the SHAP summary diagram of the OSC model are TOC, T max , pyrite, clay minerals, porosity, pyrite.

[0079] A partial dependence plot (PDP) can represent the impact of two features on the model output under interaction. Figure 8 The partial dependence results of each feature of the optimal OMC and OSC models are shown respectively.

[0080] TOC and T max They show positive and negative contributions to the overall model output, respectively, which is also one of the most important influencing factors of OSC and OMC. This trend can be observed in both OMC and OSC models. In addition, pyrite shows a positive contribution to both OMC and OSC. The interaction between clay minerals and quartz shows the influence of mineral components on OMC and OSC predictions. Among them, clay minerals and quartz contribute very little to OMC, and the overall impact is not significant. However, in the OSC model, the clay mineral content shows a positive contribution to the model, while the increase in calcite content has a negative effect on OSC. Finally, porosity shows a positive contribution to both OMC and OSC. It is worth noting that the macropore volume has a significant positive contribution to the model in OMC prediction, but its contribution is not prominent in OSC prediction.

[0081] As an important geological feature of shale, pore structure parameters have a significant impact on shale OMC and OSC, which is why these parameters are used in the present invention to predict OMC and OSC. However, in actual petroleum geological research, due to the high cost of obtaining pore structure parameters, there may be a problem of missing pore structure parameter data. In order to highlight the advantages of the present invention, here we only select mineralogical parameters and basic geochemical parameters, use the same method to build the model, and compare the performance of the OMC and OSC prediction models with and without considering pore structure parameters.

[0082] Experiments were conducted using a tree model with relatively stable prediction performance ( Fig. 9 The results show that the prediction performance of the prediction model considering pore structure parameters is better than that of the prediction model not considering pore structure parameters. The average prediction R 2 The average predicted R values ​​of the optimal OMC and OSC models without considering the pore structure parameters are 0.82 and 0.77, respectively, and the average RMSE values ​​are 0.24 and 0.20, respectively. 2 The average RMSEs are 0.26 and 0.21, respectively. In addition to the difference in the performance of the optimal prediction model, the average prediction performance of the prediction model considering the pore structure parameters is more stable than that of the prediction model without considering the pore structure parameters, which indicates that the addition of detailed pore structure parameter information improves the prediction performance of the model, which is conducive to achieving more accurate and stable OMC and OSC predictions.

[0083] The specific implementation examples described in this article are merely examples of the spirit of the present invention. Those skilled in the art of the present invention may make various modifications or supplements to the specific implementation examples described or take corresponding measures, but they will not deviate from the spirit of the present invention or exceed the scope defined by the attached claims.

Claims

1. A method for predicting the mobility and adsorption of shale oil based on pore structure parameters, characterized in that: The following steps are involved: S1, using mineralogy and basic geochemical parameters, the pore structure parameters are supplemented by Missforest iterative interpolation to obtain pore structure parameter data; S2, the original data set is composed of mineralogy type, basic geochemical parameters, porosity, pore structure parameter data and multi-temperature pyrolysis data; S3, inputting the original data set into the prediction model to obtain the prediction results of shale oil mobility and adsorption.

2. The method for predicting shale oil mobility and adsorption based on pore structure parameters according to claim 1, characterized in that: The mineralogical types include quartz, feldspar, plagioclase, potassium feldspar, calcite, dolomite, aragonite, siderite, gypsum, anhydrite, clay minerals and pyrite; the basic geochemical parameters include organic matter abundance and pyrolysis hydrocarbon peak temperature.

3. The method for predicting shale oil mobility and adsorption based on pore structure parameters according to claim 1, characterized in that: The pore structure parameter data include the volume and specific surface area of ​​different types of pores, wherein the pore types include micropores, transitional pores, mesopores and macropores.

4. The method for predicting shale oil mobility and adsorption based on pore structure parameters according to claim 1, characterized in that: During the training of the prediction model, the calculation formulas for shale oil mobility and adsorption are: Among them, OMC is the mobility of shale oil, OSC is the adsorption of shale oil, TOC is the abundance of organic matter, S 1-1 It is a multi-stage pyrolysis light oil component, S 1-2 It is a multi-stage pyrolysis light medium oil component, S 2-1 It is a multi-temperature stage pyrolysis of heavy hydrocarbons and colloidal asphalt components.

5. The method for predicting shale oil mobility and adsorption based on pore structure parameters according to claim 1, characterized in that: During the prediction model training process, the regression model is used for evaluation. The coefficient of determination and root mean square error used in the evaluation are defined as follows: In the formula, is the forecast data, is the actual value, is the actual data mean, n is the number of each data set, R 2 is the coefficient of determination and RMSE is the root mean square error.

6. The method for predicting shale oil mobility and adsorption based on pore structure parameters according to claim 1, characterized in that: The prediction models include MLP-SGD, MLP-Lbfgs, MLP-Adam, RF-PSO, RF-GA, RF-DE, XGBoost-PSO, XGBoost-GA, XGBoost-DE, SVM-PSO, SVM-GA, and SVM-DE.

7. The method for predicting shale oil mobility and adsorption based on pore structure parameters according to claim 6, characterized in that: The preferred models among the prediction models include RF-PSO and RF-PSO.

8. The method for predicting shale oil mobility and adsorption based on pore structure parameters according to any one of claims 1 to 7, characterized in that: The method also includes step S4: interpreting the prediction model using a variety of interpretability algorithms to obtain the main influencing factors of shale oil mobility and adsorption.

9. The method for predicting shale oil mobility and adsorption based on pore structure parameters according to claim 8, characterized in that: The explainability algorithms include PFI, SHAP, and PDP.

10. A device for predicting the mobility and adsorption of shale oil based on pore structure parameters, characterized in that: It includes at least one processor and a memory communicatively connected to the at least one processor; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method for predicting the mobility and adsorption of shale oil based on pore structure parameters as described in any one of claims 1 to 9.

Images